The ability of small molecules to enter cells is a critical parameter in development of pharmaceutical agents and tools for chemical biology. Most synthetic compounds enter cells by passive diffusion across the membrane. The goal of this proposal is to increase our understanding of, and ability to predict, passive membrane permeation. The ultimate test of this understanding will be rationally modifying compounds to improve membrane permeation. The work proposed here builds on existing models of membrane permeation, but the model we are developing differs from most in practical use by being more directly based on an understanding of the physics of passive membrane permeation. As such, it is systematically improvable, and has broad applicability. We propose to 1. Implement and test a physics-based model for passive membrane permeability. A first generation of this model has been described in a series of papers by the PI, and has emphasized the role of conformational flexibility and the ability to form internal hydrogen bonds in promoting membrane permeation. We will extend this model to include other critical aspects of the physics, including entropic losses upon membrane insertion and the semi-ordered hydrophobic environment of the membrane interior, comparing to both literature data and new data generated in Aims 2 and 3. 2. Interrogate key aspects of membrane permeation using cyclic peptides, and use this knowledge to design highly permeable cyclic peptides. We propose to use cyclic peptides as a challenging model system for studying passive membrane permeation. The relative synthetic ease of creating cyclic peptides facilitates developing series of compounds that differ in well-defined ways, such as stereochemistry, rigidity, size, hydrophobicity, etc. As in our earlier work, computational predictions will always be made prior to experimental testing, and the results will probe specific aspects of the physics of membrane permeation. 3. Interrogate key aspects of membrane permeation using non-peptidic small molecules, and use this knowledge in practical efforts to optimize the chemical properties of protein inhibitors. One of the collaborative projects involves improving membrane permeability for inhibitors of parasite cysteine proteases, which have typically had poor bioavailability. A central element of this proposal is collaborations with chemists Scott Lokey (UCSC) and Adam Renslo (UCSF) to test and apply the computational models. Tight coupling between computational modeling and experimental testing is central to this proposal, allowing iterative improvement of our physical understanding of passive membrane permeability and computational methods that encapsulate that understanding. Success of this work will enable hypothesis-driven ("engineering") approaches to improving membrane permeability. PUBLIC HEALTH RELEVANCE: One important property of drugs is their ability to enter cells, especially for drugs that are taken orally. This proposal is concerned with developing new computer programs that can predict the ability of compounds to enter cells, and experimental testing of these methods. Success of this work has the potential to reduce the time and cost of early-stage drug discovery, such as the proposed project to develop improved drug candidates for "sleeping sickness", a serious parasitic disease common in Africa.